Data visualization has become an integral part of business intelligence systems, making reports more informative and visually appealing and helping to fuel the development of user-friendly executive dashboards and mobile BI applications. But successful visualization efforts hinge partly on selecting the right tools for the job at hand -- and that requires more than a quick look at some colorful graphics.
One of the most elemental issues to consider is whether to buy standalone data visualization software or go with visualization tools that are offered as part of broader BI platforms. To answer that question, an evaluation and selection team will first need to acquire a good understanding of the kinds of data visualizations that business users want to see and the level of visualization functionality that's required.
Rick Sherman, founder of BI and data warehousing consultancy Athena IT Solutions in Maynard, Massachusetts, said most BI suites now provide at least basic data visualization capabilities -- for example, the ability to create simple bar and pie charts. But if an organization needs advanced features, such as support for mapping geospatial data, more specialized data discovery tools or visualization-specific software might be called for, Sherman said.
Who to turn to for visualizing data
With most companies already having BI and analytics tools in place, a related consideration is whether to stick with an existing vendor or turn to a new one. There can be obvious advantages to dealing with one vendor, or a limited number of them. Sherman, though, said those advantages might not be so great if a current vendor acquired its visualization software and hasn't fully integrated the technology into its primary BI platform. And he added that efforts to limit vendor counts are complicated by the situation on the ground in many organizations, with business units often able to make their own BI buying decisions for departmental uses.
Sherman and William McKnight, president of McKnight Consulting Group LLC in Plano, Texas, both recommended that IT or BI managers be involved in the process of evaluating and selecting data visualization tools, partly to ensure that back-end BI systems can handle the visualization load. But Sherman said a spectrum of business users should have a say as well, from the executives and operational workers who will base business decisions on visualized data to the power-user analysts who are likely to handle the job of creating visualizations. That task typically falls to "a data-savvy business person, not an IT or BI developer," he said.
In organizations that have adopted self-service BI technologies and processes, a broader set of business users might take part in building data visualizations, McKnight said. To empower such users and avoid overwhelming them, he suggested looking for data visualization software that "your mother would be able to use."
What's on the inside counts with visualization tools
Specific features that McKnight said technology evaluators should check for include the ability to drill down into visualized data for deeper analysis, support for displaying visualizations on a variety of mobile devices through HTML5 interfaces, and functionality for creating treemaps, bubble charts, infographics and other types of visualizations "that aren't your basic line or bar charts." Being able to easily map geographic data is also a useful capability, he added.
More on managing and using data visualization software
Find real-world examples and advice in our guide to advanced data visualization
Get guidance on best practices for visualizing data from consultant Lee Feinberg
More from Lee Feinberg: Fight fears about using data visualization technology
Sherman pointed to support for incorporating data visualizations into BI dashboards as another key feature -- one that in some tools "isn't as robust as you'd think it would be," he said. The same goes for data drilldown capabilities, cautioned Sherman, who also cited the ability to easily change visualization types and switch to table-based representations of data as worthwhile features to explore while evaluating visualization tools.
Connectivity to a wide variety of potential data sources, including Hadoop, NoSQL databases and cloud-based systems, is an important consideration as well, according to McKnight. That was seconded by David Loshin, president of consultancy Knowledge Integrity Inc. in Silver Spring, Maryland. "Vendors are capable of nice presentations [about their tools]," Loshin said. "But if users aren't able to process or get access to information, they aren't going to be able to get the kind of results they're looking for."